Robust Data Predictive Control Framework for Smart Multi-Microgrid Energy Dispatch Considering Electricity Market Uncertainty

Ibrahim Brahmia,Jingcheng Wang,Haotian Xu,Hongyuan Wang,Luca De Oliveira Turci

DOI: https://doi.org/10.1109/access.2021.3060315

IF: 3.9

2021-01-01

IEEE Access

Abstract:With the emerging technologies for Energy Intent (EI) and data-driven applications, the conventional power grid network is undergoing a radical modernization. An efficient energy management and electricity price forecasting remains a challenging task. In this paper, a new Robust Data Predictive Control framework for Energy Management System (RDPC-EMS) is developed to overcome the uncertainty of the electricity retail price market and minimize the total operating costs for the multi-microgrids (MMG) system. The proposed framework solves the economic energy dispatch based on an accurate Electricity Price Forecasting (EPF) by an Outlier-Robust Extreme Learning Machine (OR-ELM) algorithm and a two layers cooperative Distributed Model Predictive Control (DMPC). The First level provides an optimal energy scheduling between the Distribution System Operator (DSO) and cooperative microgrids systems to minimize the operating cost based on the forecasted electricity price. In contrast, second level maintains the supply-demand balance by applying the optimal energy scheduling from the first layer through an adjustment of the distributed energy resources (DER). The electricity retail price prediction is assessed using real dataset from the Iso New England electricity market. The OR-ELM regression method shows a significant forecasting performance in terms of error metrics. For instance, the mean absolute error in the training stage 2.05% for OR-ELM with a comparison of 4.17% and 6.29% for Support Vector Regression (SVR), and Artificial Neural Network (ANN) models respectively. Finally, simulation results demonstrate the efficiency RDPC-EMS for daily operating cost reduction, with decrease of 15% for MG 1 and 16% for MG 2.